Dockerfile

# Base image with GPU support
FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04

# Set env vars
ENV DEBIAN_FRONTEND=noninteractive
ENV PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1

# Install Python and other system dependencies
RUN apt-get update && apt-get install -y \
    python3 python3-pip ffmpeg curl git wget sudo \
    && rm -rf /var/lib/apt/lists/*

# Ensure `python` points to `python3`
RUN ln -sf /usr/bin/python3 /usr/bin/python && pip install --upgrade pip

# Add a user to avoid running everything as root (best practice)
RUN useradd -ms /bin/bash ollama
USER ollama
WORKDIR /home/ollama

# Install Ollama as user
RUN curl -fsSL https://ollama.com/install.sh | sh

# Switch back to root to install Python packages globally
USER root

# Create working directory for your app
WORKDIR /app

# Copy project files
COPY . .

# Download SpeechBrain model interface if needed
RUN wget -O src/custome_interface.py https://huggingface.co/Jzuluaga/accent-id-commonaccent_xlsr-en-english/resolve/main/custom_interface.py

# Install Python dependencies
RUN pip install -r requirements.txt

# Expose Streamlit default port
EXPOSE 8501

# Entrypoint: Start Ollama server, wait a bit, pull model, then launch Streamlit
CMD bash -c "\
    /home/ollama/.ollama/bin/ollama serve & \
    sleep 5 && \
    /home/ollama/.ollama/bin/ollama pull gemma3 && \
    streamlit run streamlit_app.py --server.port=8501 --server.address=0.0.0.0"